LIP6 1999/015:
THÈSE de DOCTORAT de l'UNIVERSITÉ PARIS 6 LIP6 /
LIP6
research reports
268 pages - Juin/June 1999 -
French document.
PostScript : 3054 Ko /Kb
Contact : par mail / e-mail
Thème/Team: Apprentissage et Acquisition de Connaissances
Titre français : S.A.G.A.C.E.
Solution Algorithmique Génétique pour lAnticipation de Comportements Evolutifs Application aux jeux à information complète et imparfaite
Titre anglais : S.A.G.A.C.E.
Genetic Algorithmic Solution for Evolving Behaviors' Anticipation Application to games with complete and imperfect information
Abstract : In a world where men increasingly interact with hardware and/or software, it is advisable to give machines the means to adapt to their users. The machine should serve Man and not the reverse. In order to maximize efficiency, it has to be autonomous and capable of anticipating the users needs, desires and behaviors. The task is complicated by the fact that Man is not a machine himself : he is adaptive, changes his mind, often acts irrationally
The work that is described in this thesis lies within this framework; It concerns the development of a modular method for anticipating changing (adaptive) behaviors : S.A.G.A.C.E.
Games allow machines to confront humans in a context where they must make the most of their abilities of analysis, reflection and adaptation. The implementation of a system capable of anticipating human strategies must allow, in the long run, to manage the anticipation of human behavior in many other activities and thus to fulfil, at least partially, the above mentioned objective.
Key-words : Anticipation, Learning, Games, Human player modelling, Game theory
Publications internes LIP6 1999 / LIP6 research reports 1999